Original article
Left atrial function in cardiac amyloidosis Gianluca Di Bellaa, Fabio Minutolib, Antonio Madaffaria, Anna Mazzeoc, Massimo Russoc, Rocco Donatob, Concetta Zitoa, Giovanni D. Aquarod, Maurizio Cusma` Piccionea, Stefano Pedrif, Giuseppe Vitac, Alessandro Pingitoree and Scipione Carerja Aims Left atrium can be involved by amyloid deposition in familial amyloid polyneuropathy (FAP). The aim of our study is to assess left atrium function in atrial amyloidosis. Methods Twenty-eight FAP patients (53 W 12 years) and a control group of 22 asymptomatic individuals (49 W 11 years) underwent strain echocardiography and cardiac magnetic resonance (CMR). CMR by late gadolinium enhancement (LGE) was used to assess the left atrium amyloid deposition, whereas strain echocardiography was used to quantify the left atrium deformation. The following atrial longitudinal strain (ALS) parameters were assessed: peak at the end of ventricular systole (peak-ALS), peak at early diastole (early-ALS), negative peak in late diastole, precontraction (prec)-ALS (difference between peak-ALS and early-ALS), and late ALS (sum of negative peak and prec-ALS). Results CMR showed atrial LGE in 14 FAP patients (LGE-atrial group), whereas 14 FAP patients showed no LGE (no-LGE-atrial group). Peak-ALS was significantly lower in the LGE-atrial group (22.8 W 13%) compared with the no-LGE-atrial group (59.6 W 33.1%; P U 0.001) and controls (47.4 W 16.4%; P U 0.001). Early-ALS was lower in the LGE-atrial group (10.2 W 6.2%) compared with the controls (26.3 W 11.9%; P U 0.02) and the no-LGE-atrial group (30.2 W 22.4%; P U 0.01). Prec-ALS was lower
Introduction Left atrium size and function are emerging as crucial keys in many cardiovascular diseases.1–5 Left atrium function plays both a mechanical role influencing left ventricular diastolic and systolic phases, and a neuro-hormonal role contributing to release of natriuretic peptides in response to stretch.3,6 The role of the left atrium is confirmed by an adverse prognostic significance of atrial enlargement both in cardiomyopathies and in coronary artery disease.7,8
(P U 0.001) in the LGE-atrial group (12.6 W 7.8%) compared with the no-LGE-atrial group (26.2 W 15%). Conversely, late-ALS was higher (P U 0.04) in the no-LGE-atrial group (22.8 W 12.3%) compared with the controls (13.9 W 9%); no significant differences were found in the negative peak among groups. Conclusions Patients with atrial amyloidosis have an adverse left atrium remodeling associated with left atrium dysfunction. Left atrium assessment may provide useful information in the clinical and prognostic stratification of amyloidotic patients. J Cardiovasc Med 2016, 17:113–121 Keywords: 2D strain echocardiography, atrial deformation, cardiac amyloidosis, hypertrophic cardiomyopathy a Clinical and Experimental Department of Medicine, bDepartment of Biomedical Sciences and of Morphologic and Functional Images, cDepartment of Neurosciences, Psychiatry and Anaesthesiology, University of Messina, Messina, d Department of Cardiac MRI, Fondazione CNR-Regione Toscana ‘G. Monasterio,’ Pisa, eInstitute of Clinical Physiology, CNR Pisa and fEsaote, Florence, Italy
Correspondence to Gianluca Di Bella, MD, PhD, Clinical and Experimental Department of Medicine, University of Messina, Via Consolare Valeria N81, 98100 Messina, Italy Tel/fax: +39 0902213531; e-mail: [email protected] Received 22 February 2014 Revised 27 June 2014 Accepted 30 June 2014
systemic deposition of amyloidogenic variants of the transthyretin (TTR) protein, especially in the peripheral nervous system, kidneys, and heart. In this type of amyloidosis, cardiac involvement is found in at least 5–23% of patients and the left atrium is usually involved by TTR protein deposition.9,10
The mechanical function of the left atrium consists of three phases: a reservoir phase representing a passive filling during left ventricular systole; a conduit phase consisting of a passive flow of blood from the pulmonary veins during early ventricular diastole; and an active emptying – booster pump – during late ventricular diastole.1
Nowadays – thanks to the developments in noninvasive cardiac imaging – left atrium morphology and function can be accurately investigated.11–22 Namely, late gadolinium enhancement (LGE) cardiac magnetic resonance (CMR) shows the deposition of amyloid in ventricular and atrial walls,12–17 whereas strain echocardiography derived by tissue Doppler imaging and two-dimensional (2D) images permits to accurately quantify mechanical phases of ventricular and atrial deformation.1,18–20
Familial amyloid polyneuropathy (FAP) is a neurodegenerative autosomal dominant disease characterized by the
The aim of our study is to analyze atrial function in patients with and without atrial amyloid deposition.
1558-2027 ß 2016 Italian Federation of Cardiology. All rights reserved.
DOI:10.2459/JCM.0000000000000188
© 2016 Italian Federation of Cardiology. All rights reserved
114 Journal of Cardiovascular Medicine 2016, Vol 17 No 2
Methods Patients
Between September 2007 and May 2011, we enrolled 28 patients (53 12 years, 12 women) with TTR-related FAP and without any other known heart disease (arterial hypertension, significant valve disease, previous myocardial infarction, significant atrial and/or ventricular arrhythmia). The diagnosis of FAP was made at our Department of Neurosciences and was based on the following results of genetic testing for TTR: Glu89Gln mutation in 18 patients, Phe64Leu in 6 patients, Thr49Ala in 3 patients and Glu89Gln þ Gly6Ser in 1 patient. All FAP patients underwent in the same day the following: clinical evaluation to assess polyneuropathy (peripheral polyneuropathy and eventual autonomic nervous system alterations), New York Heart Association (NYHA) class, and duration of symptoms (years); an ECG evaluation to assess low voltage, Q waves, A-V block, left bundle block, The N-terminal of the prohormone brain natriuretic peptide (NT-proBNP) level dosage, and LGE-CMR; and conventional and strain echocardiography examinations. Furthermore, a control group of 22 asymptomatic individuals (49 11 years, 11 women), without cardiovascular risk factors and previous heart diseases and with normal ECG and echocardiogram, underwent strain echocardiography. Our local ethics review committee approved the study and the investigation conformed to the principles outlined in the Declaration of Helsinki. Written informed consent was obtained by all the participants. Cardiac magnetic resonance data acquisition and analysis
Cardiac magnetic resonance was performed with a 1.5-T system (Gyroscan NT; Philips Medical Systems, Best, The Netherlands) with a cardiac phased-array coil and vectorcardiogram synchronization. A breath-hold balanced fast field-echo sequence was used to evaluate left ventricular volumes, left ventricular ejection fraction (LVEF) and left ventricular mass. The sequence parameters were Time Repetition (TR)/Time Echo (TE) 3.8/1.92; flip angle 608; slice thickness 8 mm; matrix size 192 512; field of view (FOV) 300 mm, rectangular FOV 80%; number of phases 30. In each patient, depending on left ventricular volume, a total of 9–13 short-axis views and two long-axis views (fourchamber view and two-chamber view) were acquired. Late gadolinium-enhanced images in the same short-axis and long-axis views as the unenhanced images were obtained with a 2D gradient-echo inversion recovery sequence after bolus injection of 0.2 mmol/kg of gadobutrol (Gadovist, Bayer Schering Pharma, Berlin, Germany).
All LGE images analyzed in this study were obtained about 20 min after contrast medium injection with the exception of LGE images from only one patient, who required medical assistance shortly after contrast medium injection, for which only images obtained 4 min after contrast medium injection were available. The sequence parameters were TR/TE 4.3/1.54; flip angle 158; slice thickness 10 mm; matrix size 208 512; FOV 350 mm, rectangular FOV 80%. As far as inversion time is concerned, because it can be difficult to determine the optimal inversion time for nulling normal myocardium in patients with cardiac amyloidosis, we acquired multiple images in the same view using different inversion times (80–350 ms with a 30-ms increment).12,13,14 When the best inversion time for nulling the signal from normal myocardium was obtained, a total of 8–12 shortaxis views (depending on left ventricular volume) and three long-axis views (four-chamber view, two-chamber view, and left ventricular outflow view) were acquired. For data analysis, all images were transferred to and reviewed at an offline workstation. Left ventricular volumes, left ventricular mass, and LVEF were measured by an experienced cardiologist using previously validated software (EasyVision version 4.0, Philips Healthcare). All contrast-enhanced images were analyzed in consensus by a cardiologist and a radiologist each with 8 years of experience in CMR imaging. Both readers were unaware of clinical and ECG data, NT-proBNP levels, and results of echocardiography. Contrast-enhanced images were categorized as positive or negative on the basis of presence or absence of LGE abnormalities involving the left atrium (Fig. 1). Furthermore, as previously described, the presence of left ventricular LGE was also classified as subendocardial circumferential (LGE distribution over the entire subendocardial circumference, extending to various degrees into the neighboring myocardium, which was not entirely involved), focal (one or more circumscribed areas of myocardial contrast enhancement observed in the context of normal – hypointense – myocardium), or diffuse (a diminished difference in signal intensity between myocardium and blood pool).12,13 Echocardiographic data acquisition and analysis Standard two-dimensional and Doppler imaging
Echocardiographic images were obtained using a commercial ultrasound machine (My Lab 50 gold, Esaote, Florence, Italy), equipped with a 2.5-MHz phasedarray transducer. Standard 2D and Doppler echocardiographic measurements were determined in accordance with the current American Society of Echocardiography guidelines.23 Left ventricular volumes and LVEF were assessed by a modified biplane Simpson’s method.
© 2016 Italian Federation of Cardiology. All rights reserved
Atrial function in amyloidosis Di Bella et al. 115
Fig. 1
Imaging of cardiac amyloidosis in a patient with familial amyloid polyneuropathy. Echocardiographic image (a) shows symmetric left ventricular wall thickening and hyperechogenicity of the left atrium wall. Cine-CMR (b) shows atrial thickening (arrows); on LGE-CMR image (c), homogeneous high signal intensity of the entire left ventricle wall and enhancement of both right and left atria (arrows) can be seen. CMR, cardiac magnetic resonance; LGE, late gadolinium-enhanced.
The mitral peak velocity of early filling (E wave), atrial filling (A wave), E-wave deceleration time, and A-wave duration were estimated by pulsed wave Doppler. Left ventricular diastolic function was estimated by the ratio of E wave to early diastolic mitral annular velocity (E’ wave by tissue Doppler imaging), using the average early diastolic mitral annular velocities of the septal and lateral left ventricular annulus.23,24 Furthermore, longitudinal left ventricular function was obtained using S wave by tissue Doppler imaging.23 Speckle-tracking echocardiographic data
A dedicated software package (XStrain, Esaote) was used for an off-line quantification of left atrium volumes, and longitudinal left ventricular and left atrium strain. Goodquality and adequate frame rate (50–70 frames/s) images were used only. The XStrain software provides an angle-independent 2D strain based on two-dimensional echocardiographic images.25,26 More detailed, it relies on a ‘feature tracking’ algorithm, which, in order to improve the border tracking results, combines speckle tracking with other information such as tissue-to-blood border detection, periodicity of the cardiac cycle, and the fact that the cardiac borders maintain their own ‘overall spatial coherence’ over time. This software, unlike other speckle-tracking imaging software that utilize a larger region of interest (ROI),20 performs border tracking using a small ROI that is particularly useful to analyze a thin layer as the atrial wall. Particularly, left atrium border to be tracked is drawn by the operator and is identified as a sequence of points (feature tracking); frame-by-frame displacement of these points is automatically evaluated generating both volume-over-time curves and strain curves for each atrial segment. The volume-over-time curves permit automatic acquisition of the greater left atrium volume (maximal left atrium volume), the smaller left atrium volume (minimum left atrium volume), and the
measurement of the left atrium volume at the beginning of the P wave. Setting zero strain at left ventricular end diastole, the left atrium strain curve (Fig. 2a) is characterized by a positive wave that peaks at the end of ventricular systole (peak atrial longitudinal strain, peak-ALS), followed by a positive peak in early diastole (early-ALS) and a negative peak in late diastole (negative peak). Furthermore, the curve permits to identify the precontraction ALS (prec-ALS) as the difference between peak-ALS and early-ALS, and late-ALS as the sum in absolute value of negative peak and prec-ALS (Fig. 2b). Peak-ALS represents left atrium reservoir function, precALS strain and early-ALS the conduit phase, and negative peak and late-ALS the left atrium contractile function.19,27–29 As previously validated, the same software has permitted to quantify global left ventricular strain (average of 16 segments).25,26 In our study, the tracking quality was verified for each evaluated segment and subsequent manual adjustments were performed, when required. All data were elaborated with the aid of Fourier techniques that ensure a higher accuracy using the periodicity of the heart motion. The assessment of left ventricular and left atrium strain was regarded as suboptimal when tracking could not be obtained. Atrial segmental strain were obtained for 12 left atrium segments (annular, mid, and superior segments along the septal, lateral, anterior, and inferior left atrium walls using apical four-chamber view and two-chamber view images) for each patient. In each patient, a global atrial strain deformation value was obtained for each parameter by averaging the corresponding values from measurements in 12 left atrium segments. Left atrium active emptying fraction (%) was calculated from precontraction left atrium volume and minimum left atrium volume [(precontraction left atrium volume minimum left atrium
© 2016 Italian Federation of Cardiology. All rights reserved
116 Journal of Cardiovascular Medicine 2016, Vol 17 No 2
Fig. 2
(a)
(b) Peak-ALS
AL-Slate
Negative peak
Mitral inflow E
A
81
Atrial longitudinal strain
Early-ALS
Precontraction strain
Atrial longitudinal strain
Peak-ALS
54 EarlyALS
27
Negative peak
0 0
200
400
600
800
1000
1200 1400
160
790 ms
Reservoir
Conduit
Contraction
BAS SEP
MID SEP
APIC SEP
APIC LAT
MID LAT
(a) Measurement of peak atrial longitudinal strain, early atrial longitudinal strain, and negative peak from an apical four-chamber view in a normal individual. The color curves represent ALS of each segment along the cardiac cycle, whereas the dashed curve represents the average ALS. (b) Left atrium strain curve. Scheme of atrial longitudinal strain phases during cardiac cycle and their relationship with mitral inflow, ECG and left atrium phases. ALS, atrial longitudinal strain.
volume)/precontraction left atrium volume]8; delta left atrium volume was obtained using the following formula: [(maximum left atrium area minimum left atrium area)/ maximum left atrium area] 100. All images used for the analysis of left ventricular and left atrium strain were analyzed off-line by a cardiologist (G.D.B.) whose results were used for statistical analysis. Moreover, peak-ALS of 12 randomly chosen patients (4 FAP patients with atrial LGE, 4 FAP patients without atrial LGE, and 4 controls) were evaluated by a second cardiologist (A.M.), in order to evaluate interobserver variability, and by G.D.B., 3 months after the first evaluation, to determine intraobserver variability. Statistical analysis
Kolmogorov–Smirnov analysis was used to test variables for normal distribution. Continuous variables with normal distribution were expressed as means 1 SD, whereas continuous variables without normal distribution were expressed as median and interquartile ranges. Categorical variables were expressed as percentages. Analysis of variance (ANOVA) followed by Bonferroni’s post-hoc comparison tests were used to compare multiple quantitative variables. The correlation between continuous variables was tested with Pearson’s correlation coefficient. Receiver-operating characteristic (ROC) curves were constructed, and areas
under the curves were measured to determine cut-off values of strain for optimal sensitivity and specificity. Multinomial logistic regression analysis was performed to identify independent predictors of atrial LGE on cardiac MRI. For variables found to be significant in univariate analysis [peak-ALS (PALS), early-ALS, maximal/minimal atrial volumes, and left atrium active emptying fraction], we selected for multivariate analysis a parameter of atrial strain (PALS) and a parameter of atrial dimension/function (left atrium active emptying fraction). Interobserver and intraobserver variability were assessed using the Bland-Altman method. For any statistical comparison, a P value less than 0.05 was considered to be significant. Statistical analyses were performed using SPSS version 12 (SPSS Inc., Chicago, Illinois, USA) and MedCalc 6.00.014 (MedCalc Software, Mariakerke, Belgium).
Results Cardiac magnetic resonance findings
Cardiac magnetic resonance showed no left atrium LGE in 14 patients (50%) (no-LGE-atrial group) and left atrium LGE in 14 patients (50%) (LGE-atrial group). All patients of the LGE-atrial group also showed LGE at the level of the left ventricle (LV), whereas no patient of the no-LGE-atrial group showed LGE at the level of the LV.
© 2016 Italian Federation of Cardiology. All rights reserved
Atrial function in amyloidosis Di Bella et al. 117
The left ventricular LGE pattern was focal in six patients, subendocardial circumferential in two patients, and diffuse in six patients. Clinical characteristics, NT-proBNP levels, and conventional two-dimensional echocardiographic and Doppler parameters
Among 28 FAP patients, only 2 had symptoms and signs of heart failure (NYHA class III), and both were in the LGE-atrial group. The FAP patients were followed for a mean of 19 7 months; during clinical follow-up, as the only relevant clinical events, the same two patients who had NYHA class III at the first assessment were hospitalized for heart failure. Patients of the LGE-atrial group had a higher incidence of abnormal ECG findings, polyneuropathy, and duration of symptoms with respect to the control group (Table 1). The NT-proBNP level was higher in the LGE-atrial group than in the no-LGE-atrial group. Moreover, NTproBNP levels correlated with the left ventricular parameters (septal thickness: r ¼ 0.465, P < 0.05; enddiastolic volume: r ¼ 0.455, P < 0.05) and the left atrium parameters (minimal left atrium volume: r ¼ 0.400, P < 0.05; left atrium active emptying fraction: r ¼ 0.350, P < 0.05). There were no other statistically significant correlations between NT-proBNP levels and conventional trans-thoracic echocardiographic parameters. As reported in Table 1, many conventional echocardiographic parameters were similar among groups; namely, LVEF was similar among groups. The LGE-atrial group showed higher left ventricular septal thickness, left Table 1
ventricular mass, longitudinal dysfunction, and diastolic dysfunction than controls and no-LGE-atrial group. Speckle-tracking echocardiographic parameters: left atrial function
Maximal, minimal and precontraction left atrium volumes were greater in the LGE-atrial group compared with the controls. Although left atrium volumes were higher in the LGE-atrial group with respect to the noLGE-atrial group, the difference was not statistically significant. Left atrium active emptying fraction was lower in the LGE-atrial group compared with the other groups; a reduced delta left atrium volume was found in the LGE-atrial group with respect to the other groups (Table 2). Atrial strain parameters have been measured in 558/600 segments (93%); the remaining segments were excluded from the analysis because of poor echocardiographic image quality or inadequate border tracking. Global peak-ALS was significantly lower in the LGEatrial group (22.8 13%) compared with the no-LGEatrial group (59.6 33.1%; P ¼ 0.002) and controls (47.4 16.4%; P ¼ 0.001), whereas it was similar between the no-LGE-atrial group and controls. Global early-ALS was lower in the LGE-atrial group (10.2 6.2%) as compared with the controls (26.3 11.9%; P < 0.0001) and the no-LGE-atrial group (30.2 22.4%; P ¼ 0.01), whereas it was similar between the no-LGE-atrial group and the controls. Global prec-ALS was lower (P ¼ 0.001) in the LGE-atrial group (12.6 7.8%) as compared with the no-LGE-atrial group (26.2 15%; P ¼ 0.01). No statistically significant differences were found in global prec-ALS between controls (16.8 6.5%) and the
Baseline, echocardiography, cardiac magnetic resonance, and NT-proBNP data
Age (years) Abnormal ECG findings (%) Polyneuropathy (%) NYHA >2 (%) Duration of symptoms (years) ECHO septal thickness (mm) ECHO – LV EDD (mm) ECHO – LV ESD (mm) ECHO – LV EDV (ml/m2) ECHO – LV ESV (ml/m2) ECHO – LVEF (%) ECHO – LV mass (g/m2) Normal E/A (%) E/E’ ratio S’ (cm/s) CMR – LV EDV (ml/m2) CMR – LV ESV (ml/m2) CMR – LV EF (%) CMR – LV mass (g/m2) NT-proBNP level (pg/ml)
Control group (22 patients)
No-LGE-atrial group (14 patients)
CA LGE-atrial (14 patients)
P
P
49 11 0 5 0 0 10 1 45.6 3.9 29.6 6 48.4 12 13.5 8 65.6 6.7 68 13 73 6.7 4 9.3 2.3 91 17 37 7.5 58 1.4 57 22 —————
51 12 0 43 0 1.6 2.2 11.5 1.9 48.2 9 28.9 6 46 9.6 15.7 6.5 66.4 8.2 84 22 46 5.7 4.6 8.4 2.2 85 12 35 12 60.2 5.9 65 23 44 (48)
56 12 57 93 14 2.3 2.9 18.4 3.3 47.1 7 32 11.7 41.4 14.7 13.4 10.9 62 13 195 56 43 13.8 8 6.3 1.4 89 29 40 16 54 7.4 93 32 466 (465)
NS NS 0.003 NS 0.03 NS NS NS NS NS NS NS NS 0.003 NS NS NS NS NS —
NS
Left atrial function in cardiac amyloidosis Gianluca Di Bellaa, Fabio Minutolib, Antonio Madaffaria, Anna Mazzeoc, Massimo Russoc, Rocco Donatob, Concetta Zitoa, Giovanni D. Aquarod, Maurizio Cusma` Piccionea, Stefano Pedrif, Giuseppe Vitac, Alessandro Pingitoree and Scipione Carerja Aims Left atrium can be involved by amyloid deposition in familial amyloid polyneuropathy (FAP). The aim of our study is to assess left atrium function in atrial amyloidosis. Methods Twenty-eight FAP patients (53 W 12 years) and a control group of 22 asymptomatic individuals (49 W 11 years) underwent strain echocardiography and cardiac magnetic resonance (CMR). CMR by late gadolinium enhancement (LGE) was used to assess the left atrium amyloid deposition, whereas strain echocardiography was used to quantify the left atrium deformation. The following atrial longitudinal strain (ALS) parameters were assessed: peak at the end of ventricular systole (peak-ALS), peak at early diastole (early-ALS), negative peak in late diastole, precontraction (prec)-ALS (difference between peak-ALS and early-ALS), and late ALS (sum of negative peak and prec-ALS). Results CMR showed atrial LGE in 14 FAP patients (LGE-atrial group), whereas 14 FAP patients showed no LGE (no-LGE-atrial group). Peak-ALS was significantly lower in the LGE-atrial group (22.8 W 13%) compared with the no-LGE-atrial group (59.6 W 33.1%; P U 0.001) and controls (47.4 W 16.4%; P U 0.001). Early-ALS was lower in the LGE-atrial group (10.2 W 6.2%) compared with the controls (26.3 W 11.9%; P U 0.02) and the no-LGE-atrial group (30.2 W 22.4%; P U 0.01). Prec-ALS was lower
Introduction Left atrium size and function are emerging as crucial keys in many cardiovascular diseases.1–5 Left atrium function plays both a mechanical role influencing left ventricular diastolic and systolic phases, and a neuro-hormonal role contributing to release of natriuretic peptides in response to stretch.3,6 The role of the left atrium is confirmed by an adverse prognostic significance of atrial enlargement both in cardiomyopathies and in coronary artery disease.7,8
(P U 0.001) in the LGE-atrial group (12.6 W 7.8%) compared with the no-LGE-atrial group (26.2 W 15%). Conversely, late-ALS was higher (P U 0.04) in the no-LGE-atrial group (22.8 W 12.3%) compared with the controls (13.9 W 9%); no significant differences were found in the negative peak among groups. Conclusions Patients with atrial amyloidosis have an adverse left atrium remodeling associated with left atrium dysfunction. Left atrium assessment may provide useful information in the clinical and prognostic stratification of amyloidotic patients. J Cardiovasc Med 2016, 17:113–121 Keywords: 2D strain echocardiography, atrial deformation, cardiac amyloidosis, hypertrophic cardiomyopathy a Clinical and Experimental Department of Medicine, bDepartment of Biomedical Sciences and of Morphologic and Functional Images, cDepartment of Neurosciences, Psychiatry and Anaesthesiology, University of Messina, Messina, d Department of Cardiac MRI, Fondazione CNR-Regione Toscana ‘G. Monasterio,’ Pisa, eInstitute of Clinical Physiology, CNR Pisa and fEsaote, Florence, Italy
Correspondence to Gianluca Di Bella, MD, PhD, Clinical and Experimental Department of Medicine, University of Messina, Via Consolare Valeria N81, 98100 Messina, Italy Tel/fax: +39 0902213531; e-mail: [email protected] Received 22 February 2014 Revised 27 June 2014 Accepted 30 June 2014
systemic deposition of amyloidogenic variants of the transthyretin (TTR) protein, especially in the peripheral nervous system, kidneys, and heart. In this type of amyloidosis, cardiac involvement is found in at least 5–23% of patients and the left atrium is usually involved by TTR protein deposition.9,10
The mechanical function of the left atrium consists of three phases: a reservoir phase representing a passive filling during left ventricular systole; a conduit phase consisting of a passive flow of blood from the pulmonary veins during early ventricular diastole; and an active emptying – booster pump – during late ventricular diastole.1
Nowadays – thanks to the developments in noninvasive cardiac imaging – left atrium morphology and function can be accurately investigated.11–22 Namely, late gadolinium enhancement (LGE) cardiac magnetic resonance (CMR) shows the deposition of amyloid in ventricular and atrial walls,12–17 whereas strain echocardiography derived by tissue Doppler imaging and two-dimensional (2D) images permits to accurately quantify mechanical phases of ventricular and atrial deformation.1,18–20
Familial amyloid polyneuropathy (FAP) is a neurodegenerative autosomal dominant disease characterized by the
The aim of our study is to analyze atrial function in patients with and without atrial amyloid deposition.
1558-2027 ß 2016 Italian Federation of Cardiology. All rights reserved.
DOI:10.2459/JCM.0000000000000188
© 2016 Italian Federation of Cardiology. All rights reserved
114 Journal of Cardiovascular Medicine 2016, Vol 17 No 2
Methods Patients
Between September 2007 and May 2011, we enrolled 28 patients (53 12 years, 12 women) with TTR-related FAP and without any other known heart disease (arterial hypertension, significant valve disease, previous myocardial infarction, significant atrial and/or ventricular arrhythmia). The diagnosis of FAP was made at our Department of Neurosciences and was based on the following results of genetic testing for TTR: Glu89Gln mutation in 18 patients, Phe64Leu in 6 patients, Thr49Ala in 3 patients and Glu89Gln þ Gly6Ser in 1 patient. All FAP patients underwent in the same day the following: clinical evaluation to assess polyneuropathy (peripheral polyneuropathy and eventual autonomic nervous system alterations), New York Heart Association (NYHA) class, and duration of symptoms (years); an ECG evaluation to assess low voltage, Q waves, A-V block, left bundle block, The N-terminal of the prohormone brain natriuretic peptide (NT-proBNP) level dosage, and LGE-CMR; and conventional and strain echocardiography examinations. Furthermore, a control group of 22 asymptomatic individuals (49 11 years, 11 women), without cardiovascular risk factors and previous heart diseases and with normal ECG and echocardiogram, underwent strain echocardiography. Our local ethics review committee approved the study and the investigation conformed to the principles outlined in the Declaration of Helsinki. Written informed consent was obtained by all the participants. Cardiac magnetic resonance data acquisition and analysis
Cardiac magnetic resonance was performed with a 1.5-T system (Gyroscan NT; Philips Medical Systems, Best, The Netherlands) with a cardiac phased-array coil and vectorcardiogram synchronization. A breath-hold balanced fast field-echo sequence was used to evaluate left ventricular volumes, left ventricular ejection fraction (LVEF) and left ventricular mass. The sequence parameters were Time Repetition (TR)/Time Echo (TE) 3.8/1.92; flip angle 608; slice thickness 8 mm; matrix size 192 512; field of view (FOV) 300 mm, rectangular FOV 80%; number of phases 30. In each patient, depending on left ventricular volume, a total of 9–13 short-axis views and two long-axis views (fourchamber view and two-chamber view) were acquired. Late gadolinium-enhanced images in the same short-axis and long-axis views as the unenhanced images were obtained with a 2D gradient-echo inversion recovery sequence after bolus injection of 0.2 mmol/kg of gadobutrol (Gadovist, Bayer Schering Pharma, Berlin, Germany).
All LGE images analyzed in this study were obtained about 20 min after contrast medium injection with the exception of LGE images from only one patient, who required medical assistance shortly after contrast medium injection, for which only images obtained 4 min after contrast medium injection were available. The sequence parameters were TR/TE 4.3/1.54; flip angle 158; slice thickness 10 mm; matrix size 208 512; FOV 350 mm, rectangular FOV 80%. As far as inversion time is concerned, because it can be difficult to determine the optimal inversion time for nulling normal myocardium in patients with cardiac amyloidosis, we acquired multiple images in the same view using different inversion times (80–350 ms with a 30-ms increment).12,13,14 When the best inversion time for nulling the signal from normal myocardium was obtained, a total of 8–12 shortaxis views (depending on left ventricular volume) and three long-axis views (four-chamber view, two-chamber view, and left ventricular outflow view) were acquired. For data analysis, all images were transferred to and reviewed at an offline workstation. Left ventricular volumes, left ventricular mass, and LVEF were measured by an experienced cardiologist using previously validated software (EasyVision version 4.0, Philips Healthcare). All contrast-enhanced images were analyzed in consensus by a cardiologist and a radiologist each with 8 years of experience in CMR imaging. Both readers were unaware of clinical and ECG data, NT-proBNP levels, and results of echocardiography. Contrast-enhanced images were categorized as positive or negative on the basis of presence or absence of LGE abnormalities involving the left atrium (Fig. 1). Furthermore, as previously described, the presence of left ventricular LGE was also classified as subendocardial circumferential (LGE distribution over the entire subendocardial circumference, extending to various degrees into the neighboring myocardium, which was not entirely involved), focal (one or more circumscribed areas of myocardial contrast enhancement observed in the context of normal – hypointense – myocardium), or diffuse (a diminished difference in signal intensity between myocardium and blood pool).12,13 Echocardiographic data acquisition and analysis Standard two-dimensional and Doppler imaging
Echocardiographic images were obtained using a commercial ultrasound machine (My Lab 50 gold, Esaote, Florence, Italy), equipped with a 2.5-MHz phasedarray transducer. Standard 2D and Doppler echocardiographic measurements were determined in accordance with the current American Society of Echocardiography guidelines.23 Left ventricular volumes and LVEF were assessed by a modified biplane Simpson’s method.
© 2016 Italian Federation of Cardiology. All rights reserved
Atrial function in amyloidosis Di Bella et al. 115
Fig. 1
Imaging of cardiac amyloidosis in a patient with familial amyloid polyneuropathy. Echocardiographic image (a) shows symmetric left ventricular wall thickening and hyperechogenicity of the left atrium wall. Cine-CMR (b) shows atrial thickening (arrows); on LGE-CMR image (c), homogeneous high signal intensity of the entire left ventricle wall and enhancement of both right and left atria (arrows) can be seen. CMR, cardiac magnetic resonance; LGE, late gadolinium-enhanced.
The mitral peak velocity of early filling (E wave), atrial filling (A wave), E-wave deceleration time, and A-wave duration were estimated by pulsed wave Doppler. Left ventricular diastolic function was estimated by the ratio of E wave to early diastolic mitral annular velocity (E’ wave by tissue Doppler imaging), using the average early diastolic mitral annular velocities of the septal and lateral left ventricular annulus.23,24 Furthermore, longitudinal left ventricular function was obtained using S wave by tissue Doppler imaging.23 Speckle-tracking echocardiographic data
A dedicated software package (XStrain, Esaote) was used for an off-line quantification of left atrium volumes, and longitudinal left ventricular and left atrium strain. Goodquality and adequate frame rate (50–70 frames/s) images were used only. The XStrain software provides an angle-independent 2D strain based on two-dimensional echocardiographic images.25,26 More detailed, it relies on a ‘feature tracking’ algorithm, which, in order to improve the border tracking results, combines speckle tracking with other information such as tissue-to-blood border detection, periodicity of the cardiac cycle, and the fact that the cardiac borders maintain their own ‘overall spatial coherence’ over time. This software, unlike other speckle-tracking imaging software that utilize a larger region of interest (ROI),20 performs border tracking using a small ROI that is particularly useful to analyze a thin layer as the atrial wall. Particularly, left atrium border to be tracked is drawn by the operator and is identified as a sequence of points (feature tracking); frame-by-frame displacement of these points is automatically evaluated generating both volume-over-time curves and strain curves for each atrial segment. The volume-over-time curves permit automatic acquisition of the greater left atrium volume (maximal left atrium volume), the smaller left atrium volume (minimum left atrium volume), and the
measurement of the left atrium volume at the beginning of the P wave. Setting zero strain at left ventricular end diastole, the left atrium strain curve (Fig. 2a) is characterized by a positive wave that peaks at the end of ventricular systole (peak atrial longitudinal strain, peak-ALS), followed by a positive peak in early diastole (early-ALS) and a negative peak in late diastole (negative peak). Furthermore, the curve permits to identify the precontraction ALS (prec-ALS) as the difference between peak-ALS and early-ALS, and late-ALS as the sum in absolute value of negative peak and prec-ALS (Fig. 2b). Peak-ALS represents left atrium reservoir function, precALS strain and early-ALS the conduit phase, and negative peak and late-ALS the left atrium contractile function.19,27–29 As previously validated, the same software has permitted to quantify global left ventricular strain (average of 16 segments).25,26 In our study, the tracking quality was verified for each evaluated segment and subsequent manual adjustments were performed, when required. All data were elaborated with the aid of Fourier techniques that ensure a higher accuracy using the periodicity of the heart motion. The assessment of left ventricular and left atrium strain was regarded as suboptimal when tracking could not be obtained. Atrial segmental strain were obtained for 12 left atrium segments (annular, mid, and superior segments along the septal, lateral, anterior, and inferior left atrium walls using apical four-chamber view and two-chamber view images) for each patient. In each patient, a global atrial strain deformation value was obtained for each parameter by averaging the corresponding values from measurements in 12 left atrium segments. Left atrium active emptying fraction (%) was calculated from precontraction left atrium volume and minimum left atrium volume [(precontraction left atrium volume minimum left atrium
© 2016 Italian Federation of Cardiology. All rights reserved
116 Journal of Cardiovascular Medicine 2016, Vol 17 No 2
Fig. 2
(a)
(b) Peak-ALS
AL-Slate
Negative peak
Mitral inflow E
A
81
Atrial longitudinal strain
Early-ALS
Precontraction strain
Atrial longitudinal strain
Peak-ALS
54 EarlyALS
27
Negative peak
0 0
200
400
600
800
1000
1200 1400
160
790 ms
Reservoir
Conduit
Contraction
BAS SEP
MID SEP
APIC SEP
APIC LAT
MID LAT
(a) Measurement of peak atrial longitudinal strain, early atrial longitudinal strain, and negative peak from an apical four-chamber view in a normal individual. The color curves represent ALS of each segment along the cardiac cycle, whereas the dashed curve represents the average ALS. (b) Left atrium strain curve. Scheme of atrial longitudinal strain phases during cardiac cycle and their relationship with mitral inflow, ECG and left atrium phases. ALS, atrial longitudinal strain.
volume)/precontraction left atrium volume]8; delta left atrium volume was obtained using the following formula: [(maximum left atrium area minimum left atrium area)/ maximum left atrium area] 100. All images used for the analysis of left ventricular and left atrium strain were analyzed off-line by a cardiologist (G.D.B.) whose results were used for statistical analysis. Moreover, peak-ALS of 12 randomly chosen patients (4 FAP patients with atrial LGE, 4 FAP patients without atrial LGE, and 4 controls) were evaluated by a second cardiologist (A.M.), in order to evaluate interobserver variability, and by G.D.B., 3 months after the first evaluation, to determine intraobserver variability. Statistical analysis
Kolmogorov–Smirnov analysis was used to test variables for normal distribution. Continuous variables with normal distribution were expressed as means 1 SD, whereas continuous variables without normal distribution were expressed as median and interquartile ranges. Categorical variables were expressed as percentages. Analysis of variance (ANOVA) followed by Bonferroni’s post-hoc comparison tests were used to compare multiple quantitative variables. The correlation between continuous variables was tested with Pearson’s correlation coefficient. Receiver-operating characteristic (ROC) curves were constructed, and areas
under the curves were measured to determine cut-off values of strain for optimal sensitivity and specificity. Multinomial logistic regression analysis was performed to identify independent predictors of atrial LGE on cardiac MRI. For variables found to be significant in univariate analysis [peak-ALS (PALS), early-ALS, maximal/minimal atrial volumes, and left atrium active emptying fraction], we selected for multivariate analysis a parameter of atrial strain (PALS) and a parameter of atrial dimension/function (left atrium active emptying fraction). Interobserver and intraobserver variability were assessed using the Bland-Altman method. For any statistical comparison, a P value less than 0.05 was considered to be significant. Statistical analyses were performed using SPSS version 12 (SPSS Inc., Chicago, Illinois, USA) and MedCalc 6.00.014 (MedCalc Software, Mariakerke, Belgium).
Results Cardiac magnetic resonance findings
Cardiac magnetic resonance showed no left atrium LGE in 14 patients (50%) (no-LGE-atrial group) and left atrium LGE in 14 patients (50%) (LGE-atrial group). All patients of the LGE-atrial group also showed LGE at the level of the left ventricle (LV), whereas no patient of the no-LGE-atrial group showed LGE at the level of the LV.
© 2016 Italian Federation of Cardiology. All rights reserved
Atrial function in amyloidosis Di Bella et al. 117
The left ventricular LGE pattern was focal in six patients, subendocardial circumferential in two patients, and diffuse in six patients. Clinical characteristics, NT-proBNP levels, and conventional two-dimensional echocardiographic and Doppler parameters
Among 28 FAP patients, only 2 had symptoms and signs of heart failure (NYHA class III), and both were in the LGE-atrial group. The FAP patients were followed for a mean of 19 7 months; during clinical follow-up, as the only relevant clinical events, the same two patients who had NYHA class III at the first assessment were hospitalized for heart failure. Patients of the LGE-atrial group had a higher incidence of abnormal ECG findings, polyneuropathy, and duration of symptoms with respect to the control group (Table 1). The NT-proBNP level was higher in the LGE-atrial group than in the no-LGE-atrial group. Moreover, NTproBNP levels correlated with the left ventricular parameters (septal thickness: r ¼ 0.465, P < 0.05; enddiastolic volume: r ¼ 0.455, P < 0.05) and the left atrium parameters (minimal left atrium volume: r ¼ 0.400, P < 0.05; left atrium active emptying fraction: r ¼ 0.350, P < 0.05). There were no other statistically significant correlations between NT-proBNP levels and conventional trans-thoracic echocardiographic parameters. As reported in Table 1, many conventional echocardiographic parameters were similar among groups; namely, LVEF was similar among groups. The LGE-atrial group showed higher left ventricular septal thickness, left Table 1
ventricular mass, longitudinal dysfunction, and diastolic dysfunction than controls and no-LGE-atrial group. Speckle-tracking echocardiographic parameters: left atrial function
Maximal, minimal and precontraction left atrium volumes were greater in the LGE-atrial group compared with the controls. Although left atrium volumes were higher in the LGE-atrial group with respect to the noLGE-atrial group, the difference was not statistically significant. Left atrium active emptying fraction was lower in the LGE-atrial group compared with the other groups; a reduced delta left atrium volume was found in the LGE-atrial group with respect to the other groups (Table 2). Atrial strain parameters have been measured in 558/600 segments (93%); the remaining segments were excluded from the analysis because of poor echocardiographic image quality or inadequate border tracking. Global peak-ALS was significantly lower in the LGEatrial group (22.8 13%) compared with the no-LGEatrial group (59.6 33.1%; P ¼ 0.002) and controls (47.4 16.4%; P ¼ 0.001), whereas it was similar between the no-LGE-atrial group and controls. Global early-ALS was lower in the LGE-atrial group (10.2 6.2%) as compared with the controls (26.3 11.9%; P < 0.0001) and the no-LGE-atrial group (30.2 22.4%; P ¼ 0.01), whereas it was similar between the no-LGE-atrial group and the controls. Global prec-ALS was lower (P ¼ 0.001) in the LGE-atrial group (12.6 7.8%) as compared with the no-LGE-atrial group (26.2 15%; P ¼ 0.01). No statistically significant differences were found in global prec-ALS between controls (16.8 6.5%) and the
Baseline, echocardiography, cardiac magnetic resonance, and NT-proBNP data
Age (years) Abnormal ECG findings (%) Polyneuropathy (%) NYHA >2 (%) Duration of symptoms (years) ECHO septal thickness (mm) ECHO – LV EDD (mm) ECHO – LV ESD (mm) ECHO – LV EDV (ml/m2) ECHO – LV ESV (ml/m2) ECHO – LVEF (%) ECHO – LV mass (g/m2) Normal E/A (%) E/E’ ratio S’ (cm/s) CMR – LV EDV (ml/m2) CMR – LV ESV (ml/m2) CMR – LV EF (%) CMR – LV mass (g/m2) NT-proBNP level (pg/ml)
Control group (22 patients)
No-LGE-atrial group (14 patients)
CA LGE-atrial (14 patients)
P
P
49 11 0 5 0 0 10 1 45.6 3.9 29.6 6 48.4 12 13.5 8 65.6 6.7 68 13 73 6.7 4 9.3 2.3 91 17 37 7.5 58 1.4 57 22 —————
51 12 0 43 0 1.6 2.2 11.5 1.9 48.2 9 28.9 6 46 9.6 15.7 6.5 66.4 8.2 84 22 46 5.7 4.6 8.4 2.2 85 12 35 12 60.2 5.9 65 23 44 (48)
56 12 57 93 14 2.3 2.9 18.4 3.3 47.1 7 32 11.7 41.4 14.7 13.4 10.9 62 13 195 56 43 13.8 8 6.3 1.4 89 29 40 16 54 7.4 93 32 466 (465)
NS NS 0.003 NS 0.03 NS NS NS NS NS NS NS NS 0.003 NS NS NS NS NS —
NS
